Today’s Internet of Things (IoT) is often employed to connect material artefacts to digital identifiers and a digital record of their history and existence. This has been heralded as a coming together of our material existences and our increasingly-digital lives. Bringing each object that we create, use and cherish into the IoT, is an outwardly appealing prospect. Using material objects is an accepted part of connecting with narratives and our history, and such a technological boon already enables the storytelling opportunities that are supported by rich digital records. However, in everyday life and in the practices that occupy them, people consider and share stories about the things that they feel to be meaningful to them in complex ways which do not necessarily conform to the expectations of the designers and developers who attempt to intervene and support such practices by focusing on the material objects at hand. This thesis draws upon observations from a thorough engagement with the community of practice of the Tabletop Miniature Wargaming pastime, which involves the acknowledged craft and use of objects deemed as meaningful, to reveal that the practitioners, in reality, construct their shared records and narratives around intangible Identities, both singular and collective, which they find to be the actual ‘meaningful things’ of their activities. These findings contravene the conventional emphasis on the material objects, and pose technological and conceptual challenges. Considering these findings through a lens informed by philosophical grounding, the thesis examines the distinctions between ordinary objects and extraordinary things; how things become meaningful; and the interplay between material and abstract things. The culmination of these efforts is the Meaningful Things Framework, which aims to help disambiguate the complex ways by which practitioners create, perceive and treat the meaningful things involved in their activities, and aid designers, developers and the communities themselves in understanding and supporting their practices.

With a worldwide shift towards a knowledge economy, universities are seen as a fundamental driver of economic growth. While previous studies have focused on universities’ more direct commercialisation activities, this research investigates the relatively unexplored influence of university education upon graduate entrepreneurship. By considering the digital economy, this exploratory study examines a fast growing sector where knowledge is considered to be a core asset. A global open dataset of digital start-ups is used to evaluate universities’ contribution to the performance of these ventures through the provision of formal technical, business or more general education. The impact of this human capital contribution on the probability of being a habitual entrepreneur in this industry sector, as well as the impact on the probability and rate of reaching the equity investment milestones of “funding” and “exit” were investigated. Prior to the data analysis, using computer science methods, unsupervised algorithms were developed to pre-process and transform the crowd-sourced dataset by linking multiple existing data sources, and it was demonstrated that this approach allows sophisticated natural language processing challenges to be overcome with relatively low technical capabilities. The consequent analysis of the transformed dataset reveals that: (1) having a founder with a university qualification significantly increases the probability of securing funding and successful exit; (2) having a founder with a university qualification in business significantly decreases the duration at which the first funding is secured and exit is achieved; (3) having a technical university qualification has no impact on the duration to securing funding, and increases the duration to exit. Following the empirical analysis, models for digital start-up teams are proposed. The thesis concludes that a consideration of the heterogeneous influence of different types of university qualifications reveals novel insights into the relationship between human capital and new venture performance.

Context: Understanding spatial patterns of biodiversity and species’ distributions is important for scientific theory, and for conservation and management of the natural world. Climatic variables are widely recognised as strong correlates of species richness over large spatial extents. Correlates of species richness at smaller extents (regional and landscape scales) are less well established, but environmental heterogeneity is widely thought to be important. A large number of environmental heterogeneity measures have been used, but in particular there is a growing interest in ‘geodiversity’, which I define here as the diversity of abiotic terrestrial and hydrological nature, comprising earth surface materials and landforms. Recent research has emphasised both geodiversity’s inherent value and its potential as a correlate and predictor of spatial biodiversity and species’ distribution patterns. However, despite this clear potential of geodiversity for improving our understanding of how patterns of life relate to environmental heterogeneity, its incorporation into biodiversity and species’ distribution modelling is substantially underdeveloped. In this thesis, using a macroecological approach I begin to address some of these knowledge gaps by analysing the relationships between geodiversity data, and its constituent ‘geofeatures’, and species richness and distributions for multiple taxa and across several scales (grain size and extent) and geographic locations. My main aims in this thesis are to more fully evaluate geodiversity itself, and improve our understanding of its role with respect to the spatial patterning of biodiversity, both conceptually and empirically. Locations and Spatial Scales Analyses were carried out within and across Great Britain (England, Scotland, and Wales) and Finland. The order of the four quantitative papers generally reflects the largest spatial extent (i.e. size of the study area) at which they were conducted, from national (PAPERS II and III) through landscape (PAPER IV), to the local scale (vegetation plots within a small upland river catchment; PAPER V). PAPER II is a study across several spatial extents (from landscape to national) and uses two grain sizes (1 km2 and 100 km2). PAPER I is a review paper that considers multiple scales and geographic locations conceptually. Time period Present day: data were from between 1995 and 2016 across all of the quantitative studies. Taxa Multiple: alien and native vascular plants across Great Britain (PAPER II); threatened bryophytes, beetles, fungi, lepidoptera, lichens, mammals, molluscs, and vascular plants across Finland (PAPER III); common and rare vascular plants across the Cairngorms, Scotland (PAPER IV); angiosperms, conifers, fungi, lichens, liverworts, lycophytes, mosses, and pteridophytes (and productivity) across an upland river catchment within the Cairngorms (PAPER V); and conceptual consideration of multiple taxa (PAPER I). Methods: For studies in Great Britain, plant data were provided by the Botanical Society of Britain and Ireland (BSBI) for PAPERS II and IV, and by the Centre for Ecology and Hydrology (CEH) for PAPER V. The threatened species data in Finland were from Finnish Environment Institute (PAPER II). Species richness (PAPERS II, III, and IV), rarity-weighted richness (RWR; PAPER III), species’ distributions (PAPERS IV and V), and productivity (measured using NDVI from colour infrared aerial imagery; PAPER V) were all analysed using Boosted Regression Tree (BRT) modelling, allowing comparisons between studies. For geodiversity data in the British studies, I compiled geodiversity data on landforms, soils, hydrological and geological features using existing national datasets (e.g. British Geological Survey), and used a geomorphometric method to extract landform coverage data (landforms included: hollows, ridges, valleys, and peaks). These data were analysed alongside environmental data, which varied between papers, relating to climate, standard topography (e.g. slope; elevation), land use, and human population. The sources of other geodiversity data in Finland, and environmental data on topography and climate, came from a variety of sources, which are detailed within each paper. Results: Geodiversity improved biodiversity and species’ distribution models throughout all of the quantitative analyses and generally declined in importance as spatial scale coarsened beyond the landscape scale. At most spatial scales and in most places, the roles of climate and/or coarse topography dominated, and geodiversity played a relatively small role, as was expected. Geodiversity, however, made consistent positive contributions to the models independently of traditionally used topographic metrics such as standard deviation of elevation and slope. Taxonomically, geodiversity: (i) was slightly more relevant for native vascular plants than alien in Great Britain (PAPER II); (ii) of similar relevance to common and rare vascular plants in the Scottish Highlands, except that the coverage of soil parent material was especially important for rare species’ distributions (PAPER IV); of similar relevance to most sessile taxa (angiosperms, fungi, mosses, liverworts, lichens, pteridophytes, and lycophytes; conifers were not related to geodiversity) in an upland Scottish river catchment (PAPER V); and more important for threatened vascular plants and bryophytes over other studied taxa in Finland (PAPER II). Geodiversity also improved models of productivity, and the variability in productivity, in PAPER V. Main conclusions and Future Directions: Geodiversity improves our understanding of, and ability to model, the relationship between biodiversity and environmental heterogeneity at multiple spatial scales, by allowing us to get closer to the real-world conditions and processes that affect life. I found that the greatest benefit comes from measuring ‘geofeatures’, which describe the constituent parts of geodiversity separately, rather than as one combined variable. Automatically extracted landform data, the use of which is novel in ecology, biogeography and macroecology, proved particularly valuable throughout this body of work, and as too did data from expert geological and hydrological maps. The idea of ‘Conserving Nature’s Stage’ (CNS), and identifying areas that are most capable of supporting high biodiversity into the future, the benefits and caveats of which are discussed in this thesis, has recently emerged. It requires a sound empirical and conceptual basis, to which my research contributes. In this thesis, I have gone some way towards demonstrating the conceptual and empirical value of incorporating geodiversity into ecological analyses across multiple spatial scales, paving the way for this recent approach to be more extensively used for theoretical and applied purposes. I accomplished this by carrying out an assessment of existing geodiversity literature and, importantly, looking forwards to consider the prospects of geodiversity within ecology (PAPER I), supported by four quantitative studies. The conservation significance is emphasised in PAPER III. Much remains to be done, however, and future research directions are detailed in PAPER I. We need to develop predictive models to test the role of geodiversity across an array of geographical and taxonomic domains, as well as to assess metrics beyond species richness and species’ distributions. One example may involve beta diversity: does spatial turnover in species relate to spatial turnover in geofeatures? Fully analysing the role of geodiversity through time will also be important, including in relation to refugia, given predicted environmental changes in climate. In progressing with this line of enquiry, we will improve our knowledge and understanding of patterns of life on Earth and, specifically, how the geophysical landscape helps shape them.

Facial Landmark Tracking (Face Tracking) is a key step for many Face Analysis systems, such as Face Recognition, Facial Expression Recognition, or Age and Gender Recognition, among others. The goal of Facial Landmark Tracking is to locate a sparse set of points defining a facial shape in a video sequence. These typically include the mouth, the eyes, the contour, or the nose tip. The state of the art method for Face Tracking builds on Cascaded Regression, in which a set of linear regressors are used in a cascaded fashion, each receiving as input the output of the previous one, subsequently reducing the error with respect to the target locations. Despite its impressive results, Cascaded Regression suffers from several drawbacks, which are basically caused by the theoretical and practical implications of using Linear Regression. Under the context of Face Alignment, Linear Regression is used to predict shape displacements from image features through a linear mapping. This linear mapping is learnt through the typical least-squares problem, in which a set of random perturbations is given. This means that, each time a new regressor is to be trained, Cascaded Regression needs to generate perturbations and apply the sampling again. Moreover, existing solutions are not capable of incorporating incremental learning in real time. It is well-known that person-specific models perform better than generic ones, and thus the possibility of personalising generic models whilst tracking is ongoing is a desired property, yet to be addressed. This thesis proposes Continuous Regression, a Functional Regression solution to the least-squares problem, resulting in the first real-time incremental face tracker. Briefly speaking, Continuous Regression approximates the samples by an estimation based on a first-order Taylor expansion yielding a closed-form solution for the infinite set of shape displacements. This way, it is possible to model the space of shape displacements as a continuum, without the need of using complex bases. Further, this thesis introduces a novel measure that allows Continuous Regression to be extended to spaces of correlated variables. This novel solution is incorporated into the Cascaded Regression framework, and its computational benefits for training under different configurations are shown. Then, it presents an approach for incremental learning within Cascaded Regression, and shows its complexity allows for real-time implementation. To the best of my knowledge, this is the first incremental face tracker that is shown to operate in real-time. The tracker is tested in an extensive benchmark, attaining state of the art results, thanks to the incremental learning capabilities.

E-books have seen a significant proliferation over recent years. In the UK, about a third of the population today owns an e-reader with about half either owning an e-reader or tablet. Nevertheless, only about 4% of readers have moved to reading e-books only. These numbers suggest that, while e-books have caught on among a large number of users, they seem to complement rather than replace books. In light of the significance of books to past and contemporary cultures and societies it is little surprising that the emergence of e-reading technologies has sparked a plethora of writing on the topic, particularly in journalism and the humanities. With a common focus on the relative merits of books and e-books, and ultimately, their respective futures (some writers go as far as either mourning or celebrating the death of the book), the debate largely suffers from a technological determinist stance, neglecting the role of social practice as a driving force in technology adoption and use. Regardless, the sheer volume of the discourse suggests that something important is at stake in the move from analogue to digital reading technologies and that books continue to be valued as physical artefacts in the digital age, if not with more fervour than ever. What is surprising then is the lack of empirical research aiming to understand how books and e-books are used and valued in everyday life. Existing work in the area is almost exclusively concerned with practices of reading, with a particular emphasis on reading in academic and professional environments, thereby not only disregarding the social and material nature of reading, but also the rich life of the book beyond its role as a reading technology. The aim of this thesis then is to provide an understanding of the practices and values surrounding books and e-books in everyday life. Based on this understanding, it further aims to explore alternatives to the current e-reading ecosystem through designs that are sensitive to some of the broader values people associate with books and e-books. To do so, it takes a situated approach to studying books and e-books as they are used over the course of their lifecycle inside and outside the home. Through a combination of a series of in-depth interviews, guided ‘home tours’, and participant diaries ‘context-rich’ data on people’s uses of, and orientations towards, books and e-books are gathered. Subsequently, design responses are iteratively developed before being returned to readers for analysis. The contribution of this thesis is fourfold: (1) an account of the socially and materially situated practices associated with books and e-books inside and outside the home, (2) an explication of the distinct, yet complementary, values reflected in and driving book and e-books use, (3) an explication of the ways in which developing a sense of self and connecting with others are actualized through the use of books and e-books, and (3) the development and in situ analysis of a design exemplar in support of these goals.

Students learning in blended learning and classroom environments benefit from social interaction and vicarious learning experience with their peers and tutors. In comparison, students learning via self-directed Web-based learning cannot benefit from these advantages and may feel isolated. This research investigates if the presence of virtual persuasive agents presented as avatars, happy images and encouraging text can provide social support similar to real peers and improve students’ Web-based learning self-efficacy (WBLSE). This research also examines if low and high fidelity virtual persuasive agents can provide social support in a similar way. This study uses Bandura’s (1982) self-efficacy theory as a framework for changing nursing students’ beliefs in using the Web for learning. The basis for including virtual persuasive agents in this research stems from the media equation theory (Reeves & Nass, 1996) which holds that Computers are Social Actors (CASA) and that people respond to Web-based media as if they were social actors. Adopting the User-Centred Design approach, a bespoke Web-based training package was developed for changing pre-registration and post-registration nursing students’ WBLSE. In a quasi-experimental design, the package was delivered in three separate studies to different groups of pre-registration and post-registration nursing students. Several important findings contributed to the WBLSE body of knowledge. Overall, the training package was found to be effective with the nursing students’ WBLSE improving equally in the intervention groups in all studies. Pre-registration students showed the greatest improvement when learning by self-direction supported by virtual persuasive agents, whereas post-registration students improved when learning in a blended setting without their support. Low-fidelity virtual persuasive agents were sufficient in providing social support for pre-registration students in self-directed settings. The implications for Web-based learning in nurse education, research and practice are discussed.

Today, the web is constantly growing, expanding global information space and more and more data is being processed and sourced online. The amount of electronically accessible and available online information is overwhelming. Increasingly, recommendation systems, which engage in some form of automated personalisation, are hugely prevalent on the web and have been extensively studied in the research literature. Several issues still remain unsolved including high sparsity situation and cold starts (how to recommend content to users who have had little or no prior interaction with the system). Recent work has demonstrated a potential solution in the form of cross-domain user modeling. This thesis will explore the design, implementation and testing of a cross-domain approach using social media data to model rich and effective user preferences and provide empirical evidence of the effectiveness of the approach based on direct real-world user feedback, deconstructing a cross-system news recommendation service where user models are generated via social media data. This will be accomplished by identifying the availability of a source domain from which to draw resources for recommendations and the availability of user profiles that capture a wide range of user interests from different domains. This thesis also demonstrates the viability of generating user models from social media data and evidences that the automated cross-domain approach can be superior to explicit filtering using self-declared preferences and can be further augmented when placing the user with the ability to maintain control over such models. The reasons for these results are qualitatively examined in order to understand why such effects occur, indicating that different models are capturing widely different areas within a user's preference space.

Advances in Biological Imaging technology have made possible imaging of sub-cellular samples with an unprecedented resolution. By using Tomographic Reconstruction biological researchers can now obtain volumetric reconstructions for whole cells in near-native state using cryo-Soft X-ray Tomography or even smaller sub-cellular regions with cryo-Electron Tomography. These technologies allow for visualisation, exploration and analysis of very exciting biological samples, however, it doesn’t come without its challenges. Poor signal-to-noise ratio, low contrast and other sample preparation and re-construction artefacts make these 3D datasets to be a great challenge for the image processing and computer vision community. Without previous available annotations due to the biological relevance of the datasets (which makes them not being publicly available) and the scarce previous research in the field, (semi-)automatic segmentation of these datasets tends to fail. In order to bring state-of-the-art in computer vision closer to the biological community and overcome the difficulties previously mentioned, we are going to build towards a semi-automatic segmentation framework. To do so, we will first introduce superpixels, a group of adjacent pixels that share similar characteristics that reduce whole images to a few superpixels that still preserve important information of the image. Superpixels have been used in the recent literature to speed up object detection, tracking and scene parsing systems. The reduced representation of the image with a few regions allows for faster processing on the subsequent algorithms applied over them. Two novel superpixel algorithms will be presented, introducing with them what we call a Super-Region Hierarchy. A Super-Region Hierarchy is composed of similar regions agglomerated hierarchically. We will show that exploiting this hierarchy in both directions (bottom-up and top-down) helps improving the quality of the superpixels and generalizing them toimages of large dimensionality. Then, superpixels are going to be extended to 3D (named supervoxels), resulting in a variation of two new algorithms ready to be applied to large biological volumes. We will show that representing biological volumes with supervoxels helps not only to dramatically reduce the computational complexity of the analysis (as billions of voxels can be accurately represented with few thousand supervoxels), but also improve the accuracy of the analysis itself by reducing the local noisy neighbourhood of these datasets when grouping voxel features within supervoxels. These regions are only as powerful as the features that represent them, and thus, an in-depth discussion about biological features and grouping methods will lead the way to our first interactive segmentation model, by gathering contextual information from super-regions and hierarchical segmentation layers to allow for segmentation of large regions of the volume with few user input (in the form of annotations or scribbles). Moving forward to improve the interactive segmentation model, a novel algorithm will be presented to extract the most representative (or relevant) sub-volumes from a 3D dataset, since the lack of training data is one of the deciding factors for automatic approaches to fail. We will show that by serving small sub-volumes to the user to be segmented and applying Active Learning to select the next best sub-volume, the number of user interactions needed to completely segment a 3D volume is dramatically reduced. A novel classifier based on Random Forests will be presented to better benefit from these regions of known shape. To finish, SuRVoS will be introduced. A novel fully functional and publicly available workbench based on the work presented here. It is a software tool that comprises most of the ideas, problem formulations and algorithms into a single user interface. It allows a user to interactively segment arbitrary volumetric datasets in a very intuitive and easy to use manner. We have then covered all the topics from data representation to segmentation of biological volumes, and provide with a software tool that hopefully will help closing the gap between biological imaging and computer vision, allowing to generate annotations (or ground truth as it is known in computer vision) much quicker with the aim of gathering a large biological segmentation database to be used in future large-scale completely automatic projects.

Human visual system has a mechanism which ensures that the perceived colour of an object remains almost constant under varying illumination conditions, and this mechanism is called colour constancy. Electronic imaging systems such as digital cameras do not naturally have this ability. The color appearance of images of an object under different lighting conditions changes with the colour of the light sources and this can cause problems in many computer vision applications such as object recognition. To deal with this problem, many algorithms have been developed to estimate the input image’s illuminant, and then recover the intrinsic colour of the scene correctly. In this thesis, we focus on this topic, try to produce new colour constancy algorithms in both images and videos, to improve the performance of the state of the art. This thesis makes four technical contributions. First, we have developed a new image representation scheme suitable for developing learning based colour constancy algorithms; second, we introduce a new method that formulates the colour constancy problem as one that infers the illuminant class of the input image; third, we introduce a novel clustering classification colour constancy framework (the 4C method); and finally, we extend our method from still image into video processing, create a new framework to deal with the colour constancy problem in videos. As in many computer vision problems, one of the crucial issues is how to effectively represent the input events. Colour constancy is no exception and we need to first represent the input image. As we are only interested in the colours of the image, colour histogram is a natural choice. However, traditional colour histogram is content dependent. As our task is estimating the colours of the illuminant rather than the colours of the image, we need a representation that is relatively independent of the image content. Based on this reasoning, we introduce the novel concept of a binary colour histogram where it records if a colour has appeared in the image or not and disregards the frequency of the colours appear in the image. We will present experimental results to demonstrate that our new binary histogram representation is particularly suitable for learning based colour constancy and that it provides better performances than other traditional representation schemes. The colour of a digital image is directly affected by the colour of the illuminant. We reason if we can recognize or classify the illuminant source of the image, we can then correct the colour of the image. Based on this rationale, we formulate the colour constancy problem as an illuminant classification problem. We assume that each image has an associated class of illuminant and the task of colour constancy is that of recognizing the illuminant class of the image. To accomplish this, we make use of our newly introduced binary colour histogram representation scheme and employ a powerful machine learning method called the Random Forest to construct the illuminant recognition system. We will present experimental results to show the effectiveness of our new method. Encouraged by the success of our illuminant recognition framework, we have developed a novel clustering classification colour constancy (the 4C) framework. We reason that similar illuminants will result in similar white point colours in an image. Based on this assumption, we first use a clustering algorithm to group similar white point colours of the training samples into the same cluster. We then treat the images in the same cluster as belonging to the same illumination source and each cluster as one class of illuminants. The colour constancy problem, i.e., that of estimating the unknown illuminant of an image, becomes that of identifying which illuminant class (cluster) the image’s illuminant falling into. We again make use of our novel binary colour histogram representation and our random forest based illuminant classification methods to implement our new 4C colour constancy framework. We present experimental results on publicly available testing datasets and show that our new method is competitive to state of the art. As a practical application, we have successfully extended our novel colour constancy methods from still image into video processing. The video tonal stabilization problem is still an unsolved problem, and current algorithms are only focusing on keeping the tonal stable during video playing, not really trying to recover the incorrect illuminant. We tackle these two problems together by keeping the tonal stable and recovering the frame colour to a canonical illuminant. Our approach first divides video frames into shots containing similar illuminant characteristics. We then correct the frames in the same scene by using the Random Forest illuminant estimation framework. A smooth function is applied to prevent flick and flash from occurring at the boundary of the neighboring scenes. Experimental results show that our new methods can improve video quality effectively.

As air traffic increases, more airports are facing capacity problems. A growing number of airports are considering optimisation methods as a solution for increasing their capacity and improving their efficiency. However, modelling an airport is complicated as there are many processes that happen in parallel, each with different constraints and objectives that need to be considered. This thesis focuses on the ground movement problem, the problem of moving aircraft efficiently around an airport. This problem links the problems at the stands and at the runways, and a good model for this problem can not only help the controllers who direct the aircraft to do so more effectively, but can also feed into taxi time estimation improvements, which can aid the solution of other optimisation problems, such as take-off sequencing. Firstly, the effects of the pushback process are investigated and a model that includes this process is presented. Secondly, the effects of different levels of prioritisation between arrivals and departures is investigated. Thirdly, the effects that the airport layout has on the routing process is investigated by examining various airport morphologies and by identifying areas that can cause delays. Different airport morphologies are compared, and the use and importance of alternative paths is highlighted. Moreover, the gate allocation process that also affects the ground movement problem is considered, and a model that uses the routing process of aircraft as a tool to provide a more informed and tailored allocation of aircraft to the gates is presented. A 52% decrease in the duration of delays was observed during the routing process of aircraft when the two processes were integrated. Furthermore, a new routing algorithm that solves the routing problem faster than what is currently used in academia for routing aircraft by taking into consideration all the available paths is presented. The results show a 46% to 67% (depending on the airport) improvement on execution time. Finally, the model is applied in a flight simulator cockpit - a tool for assisting the air transportation operations - and it was integrated with other novel technologies in other research fields. This research provides a more realistic and faster way to solve the ground movement problem of aircraft.